Results of Three Randomised Controlled Trials of Telephone Self-management

Dr Matthew Cullen (FRANZCP) Co-President, McKesson Asia-Pacific

3 September 2009

2

Agenda

1.

Chronic Disease and Self Care

2.

A Solution via Health Call Centre and Online

3.

What does the Data Reveal –

Evaluation

3

Why do we manage Chronic Disease so poorly?

50% of people with Chronic Disease have not been told about treatment options1

25% do not have Care Plan1 (Australia > 60%)2

50% do not have a Self-Care Plan1 (Australia > 75%)2

50% Medicines are not taken as intended1

1 Source: NHS 20082 Source: McK internal data

4

What is Self-Care?

A decision-making process involving…

the choice of behaviours that maintain physiological stability (self-care maintenance) and

the response to symptoms when they occur (self-care management)

Riegel, Carlson, Moser, et al 2004

5

Poor Self-Care: Worldwide Issue

Germany

(Micahelson et al, 1998)

Most common factor associated with readmission41.9% non-adherent with medicines or diet

India

(Joshi et al, 1999)

Most common cause of hospitalization (49%)Italy

(Opasisch, et al, 2001)

Most common cause of decompensation (60%)Switzerland

(Wagdi, 1993)

Common cause of decompensation (47%)31% non-adherence with medicines, 9% with sodium, 7% with fluid restriction

6

Medical Training not Chronic or Self-Care Focused

Physician preparation (feel unsure or not prepared):

Co-ordinate in-home and community services (66%)Educate patients with chronic conditions (66%)Manage the psychological and social aspects of chronic care (64%)Provide effective nutritional guidance (63%)Manage chronic pain (63%)

Source: Johns Hopkins University, Partnership for Solutions, National Public Engagement Campaign on Chronic Illness -- Physician Survey. Conducted by Mathematica Policy Research, Inc., 2001

7

Agenda

1.

Chronic Disease and Self Care

2.

A Solution via Health Call Centre and Online

3.

What does the Data Reveal –

Evaluation

Nurse coaching model in Health Call Centre

Written offer to join program (Opt in / Opt out model)

Members identified through hospital claims data (+ HRA / GP / self)

Member details transferred to health call centre

Phone based Enrollment and Assessment (RN)

Risk stratification > Determines intervention level

12 month interventionReview at 6 and 12 months

Linking in with GPs

Pre-enrolment provider mailing to inform GP of program capabilities, engage in process

Patient collaboration with GP / action plan and further enhance goal setting and awareness of risk factors

Post-enrolment mailings are patient-specific reports to keep the GP informed of the patient’s progress (collaboration)

10

Agenda

1.

Chronic Disease and Self Care

2.

A Solution via Health Call Centre and Online

3.

What does the Data Reveal –

Evaluation

11

Medibank Private RCT #1: Methods: Intention to Treat

Evaluation and study design by Monash University

Strength of randomised design is that 2 groups are comparable

Subject and controls had had a previous admission in 12 months prior with index condition

Drop outs or voluntary cross over do not occur at random

The only fair comparison is to compare the two groups as they were randomised – also called “intention to treat”

12

Medibank Private RCT #1: Methods: Statistical Analysis

T-test comparison between intervention and control groups of benefits paid, charges and costs

Matched odds ratio for self-management over time by disease (6-12 months)

Regression analysis with total cost as dependent variable across disease groups

13

RCT #1: Medibank Private – Cohort Characteristics

Not randomised by age and sex but by ID number by individual to intervention group over the same time period

Intervention Control

CHF 973 968

COPD 504 504

CAD 3882 3875

Total 5359 5347

14

Medibank Private RCT #1:Participant satisfaction survey initial results

Members were highly satisfied with staff, usefulness of calls, frequency of calls and ability to ring 24 hour advice line

70% found written action plans useful

Influence on the program on lifestyle changes was varied

▪

30% felt it helped them improve their diet

▪

57% felt it helped them take medications as prescribed

▪

63% said they had increased confidence to talk with GP

▪

70% said it increased their sense of well being

▪

77% of members wished to remain in the program

15

Medibank Private RCT #1: CHF – Change in self-management at 12 months – Not controlledSelf- management practice

Direction of change

Odds ratio Lower 95% CI Upper 95% CI p-value

Influenza vaccination

1.9 0.9 3.9 0.08

Pneumonia vaccination

- 12.8 - <0.005

Own weight scale

= 1.1 0.4 3.7 1.0

Weight record 3.8 1.8 9 <0.005

Have written action plan

22.3 8.4 83.4 <0.005

Exercise days per week

0.5 - - 0.2

16

Medibank Private RCT #1: COPD – Change in self-management at 12 months – not controlledSelf- management practice

Direction of change

Odds ratio Lower 95% CI Upper 95% CI p-value

Influenza Vaccination

3.3 1.4 9.1 <0.005

Pneumonia shot

3.1 1.2 9.7 0.015

Written action plan

7.5 4 15.7 <0.005

Passive smoking exposure

0.5 0.1 1.9 0.388

Exercise days per week

= 1.3 - - 0.277

17

Medibank Private RCT #1: CAD – Change in self-management at 12 months – not controlledSelf-management practice

Direction of change

Odds ratio Lower 95% CI Upper 95% CI p-value

Influenza vaccination

3.63 2.35 5.78 <0.01

Own weight scale 0.75 0.21 2.47 0.791

Have written action plan

14.66 10.06 22.15 <0.001

Low salt diet 4.29 2.88 6.56 <0.001

Tobacco use = 0.93 0.61 1.42 >0.05

Cholesterol check last 12 months

= 1.06 0.77 1.47 0.753

Know cholesterol results

0.66 0.5 0.87 <0.005

Know blood pressure results

1.36 1.03 1.82 0.031

Exercise days per week

1.25 0.98 1.6 0.088

18

Medibank Private RCT #1: Results for Total Benefits Paid (Bootstrap method)

Does not include McKesson chargesControl mean Intervention

meanEffect p-value

CAD 5362.60 4842.17 -520.43 0.11

CHF 9303.90 9077.86 -226.04 0.79

COPD 7877.53 7183.77 -693.77 0.39

All (adjusted)

-483.37 0.04

19

HCF RCT #2: Methods Internal analysis by HCF

The members were randomised into two cohorts using the following

variables:

Age: 30-64 years; age 65-84 years and age 75-85 yearsSex: Male and FemaleCondition: Cardiovascular, respiratory, diabetes, two of these conditions, three

or more of these conditionsSeverity: High = more admissions & days in the previous year

Low = fewer or equal admissions & days in the previous year.

Subjects were free to refuse to participate, and in the end 440 subjects were enrolled, 199 members in the McKesson cohort and 241 in the control cohort.

Paired t-tests

for data involving continuous variables, and chi-square tests for data involving all categorical variables (such as responses in the cohorts, and satisfaction surveys).

20

HCF RCT #2: Age and sex of the sample

Age 30-64 years Age 65-74 years Age 75-85 years

M F M F M FCardiovascular 202 44 119 54 81 72

Respiratory 19 22 13 20 19 29Diabetes 39 29 31 17 22 15

2 conditions 42 43 56 35 52 663 or more conditions

2 1 2

Total 304 138 219 126 175 184

21

HCF RCT #2: Statistical Methods – Cardiac subgroup: admission and benefit data

Control McKesson

Number of subjects in subgroup 123 137Percent of subjects admitted during 37 34

Percent of subjects admitted before 55 61

Mean no. admitted per subject before 1.29 1.29

Mean no. of days per subject before 5.50 5.74

ALOS before 4.25 4.45

Mean benefits paid per subject before $8229 $8868

Mean no. admitted per subject during 0.83 0.82

Mean no. of days per subject during 3.27 3.87

ALOS during 3.94 4.69

Mean benefits paid per subject during $3592 $2664

Percent benefit reduction 56 70

Improvement cf. control subgroup 24

Thus the McKesson cohort showed a greater (24%) decrease in benefits (70% cf 56%) than the control cohort. The improvement was not sustained in subjects with multiple conditions, though remained significant overall (p=0.00).

22

HCF RCT #2: (cont)

Satisfaction Levels90% felt program addressed health needs90% found the program useful70% felt they knew more about illness and health95% found information easy to understand85% followed the advice and information

SF 36

Statistically significant improvement in Intervention versus Control at 12 months

23

DVA RCT #3 – Study Design

Evaluation by University of Queensland

Veterans with Congestive Heart Failure

Combination of T – Test and Chi Squared analysis

24

DVA RCT #3 – Participants

Intervention Control

Baseline N = 214 N = 195

Per cent males 71.6 64.6

Mean age 82.3 82.5

Surveyed at 12 months N = 112 N = 181

25

DVA Study RCT #3: Cost Reduction

Reduction of $100,000 ($283,887 to $188,641) in hospital changes in treatment groupMarginal reductions in control group of $354,783 to $337,260There are very large standard deviations which makes the likelihood of being certain of statistical significanceMortality rate:

14 deaths out of 159 in intervention group (8.8%)36 deaths out of 217 in control group (16.6%)

p < 0.05 = highly clinically relevant

26

DVA RCT #3: Annual Results Congestive Heart Failure

CHF Clinical Behaviour N=Initial

Assessment12 Month

Assessment Change

Ace Inhibitor / ARB Rx 106 73% 76% ↑3

Annual Flu Vaccine 106 83% 92% ↑9

Weight Scale Availability 106 96% 93% -3

Daily Weights 99 32% 72% ↑40

Maintain Weight Log 99 15% 36% ↑21

Beta Blocker Rx 106 54% 54% -

Knows Blood Pressure 106 49% 57% ↑8

Has Action Plan 106 3% 53% ↑50

Not a Current Smoker 106 92% 97% ↑5

Low Sodium Diet 106 56% 86% ↑30

Read Labels for Na Content 106 35% 75% ↑40

Ever had Pneumonia Vaccine 106 84% 97% ↑13

27

DVA RCT #3: Chronic Heart Failure Study: Dept Veteran’s Affairs

A high level of client satisfaction and LMO acceptanceSome valuable changes in patient behaviour with respect to diet, self monitoring and medicationsAn apparent lower rate of mortality for treatment subjects (8.8% vs 16%)A possible reduction in hospital – related costs for services related to cardiovascular disease (28% reduction vs 8%)

> 1/3 study cohort aged more than 85 years

28

Conclusions

Three randomised, controlled independently conducted evaluations to dateFinancial Outcomes

Evidence of reduced hospital costs / benefit outlay

Utilisation OutcomesReduced number of admissions and length of stay

Patient OutcomesImproved condition self management eg diet, self monitoring and medicationsReduced mortality High levels of satisfaction

29

McKesson US Outcomes show positive results

Financial Outcomes Cumulative ROI of 2.22Cumulative Gross Savings: $298MCumulative Net Savings: $164M

Utilization OutcomesED visit rate decreased 25%IP Admit rate decreased 21%Rx utilization decreased 14%

Patient OutcomesOver 60% members have changed the way their manage their condition as a direct results of McKesson's DM program